Vishal Gaur, Johnson, Cornell University
Vishal Gaur, Johnson, Cornell University
Inventory Planning based on Consumer Choice in Textbook Retailing
Inventory Planning based on Consumer Choice in Textbook Retailing
07 Nov 20161:00 PM – 2:30 PM
Faculty and doctoral students only
Location:
Baker Library | Bloomberg Center 103
Organizer:
College stores compete in textbook retailing
with online stores that have significantly higher access to capital and talent.
However, a college store has data on students’ course enrollments and purchase
behavior, and could use it to make efficient stocking decisions and compete
effectively. In this paper, we describe work done with the Cornell Store for
estimating consumer choice across new and used textbooks and using it to plan
inventory in a pilot test. Our model enabled the bookstore to obtain over 10%
increase in profit in the pilot. We find that this performance is driven by two
factors. First, we are able to generate accurate demand estimates even when
nearly 90% of the textbooks in the fit sample experience stockouts. Second,
accounting for asymmetric substitution across new and used textbooks enables us
to make systematic corrections in inventory mix and inventory level compared to
the existing process. Thus, we show that it is possible to estimate demand
accurately by using an attribute-based choice model even under a high incidence
of stockouts. Further, optimal stocking rules can be distilled into simple
heuristics that can be used easily by a store manager to improve
decision-making. Finally, we discuss potential research problems in textbook
retailing arising from book rentals, returns, and used-book pricing.